On the inside back cover of books published by Gallup Press there is the following breathtaking statement:

Gallup Press exists to educate and inform the people who govern, manage, teach and lead the world's six billion citizens. Each book meets Gallup's requirements of integrity, trust and independence and is based on a Gallup-approved science and research.

Don't be distracted by the bad grammar. Focus instead on Gallup's "requirements of integrity, trust and independence." Thanks to a remarkable admission by a coauthor of Gallup's new bestseller Who Speaks for Islam? What a Billion Muslims Really Think, we are now able to know precisely what Gallup's "requirements" really are.

Who Speaks for Islam? is written by John L. Esposito, founding director of Georgetown University's Prince Alwaleed bin Talal Center for Muslim-Christian Understanding, and Dalia Mogahed, executive director of the Gallup Center for Muslim Studies. As the authors state at the outset, the book's goal is to "democratize the debate" about a potential clash between Western and Muslim civilizations by shedding light on the "actual views of everyday Muslims"--especially the "silenced majority" whose views Esposito and Mogahed argue are lost in the din about terrorism, extremism, and Islamofascism.

This majority, they contend, are just like us. They pray like Americans, dream of professional advancement like Americans, delight in technology like Americans, celebrate democracy like Americans, and cherish the ideal of women's equality like Americans. In fact, the authors write, "everyday Muslims" are so similar to ordinary Americans that "conflict between the Muslim and Western communities is far from inevitable."

Similar arguments have been made before; some of this is true, some is rubbish, much is irrelevant. The real debate about the "clash of civilizations" is about whether a determined element of radical Muslims could, like the Bolsheviks, take control of their societies and lead them into conflict with the West. The question often revolves around a disputed data point: Of the world's 1.3 billion Muslims, how many are radicals? If the number is relatively small, then the fear of a clash is inflated; if the number is relatively large, then the nightmare might not be so outlandish after all.

What gives Who Speaks for Islam? its aura of credibility is that its answers are allegedly based on hard data, not taxi-driver anecdotes from a quick visit to Cairo. The book draws on a mammoth, six-year effort to poll and interview tens of thousands of Muslims in more than 35 countries with Muslim majorities or substantial minorities. The polling sample, Esposito and Mogahed claim, represents "more than 90 percent of the world's 1.3 billion Muslims." To back up the claim, the book bears the name of the gold-standard of American polling firms, Gallup.

The answer to that all-important question, the authors say, is 7 percent. That is the percentage of Muslims who told pollsters that the attacks of September 11, 2001, were "completely" justified and who said they view the United States unfavorably--the double-barreled litmus test devised by Esposito and Mogahed to determine who is radical and who isn't.

The authors don't actually call even these people "radicals," however; the term they use is "politically radicalized," which implies that someone else is responsible for turning these otherwise ordinary Muslims into bin Laden sympathizers. By contrast, Muslims who said the 9/11 attacks were "not justified" they term "moderates."

More than half the book is an effort to distinguish the 7 percent of extremist Muslims from the "9 out of 10," as they say, who are moderates and then to focus our collective efforts on reaching out to the fringe element. With remarkable exactitude, they argue: "If the 7 percent (91 million) of the politically radicalized continue to feel politically dominated, occupied and disrespected, the West will have little, if any, chance of changing their minds." There is no need to worry about the 93 percent because, as Esposito and Mogahed have already argued, they are just like us.

There is much here to criticize. The not-so-hidden purpose of this book is to blur any difference between average Muslims around the world and average Americans, and the authors rise to the occasion at every turn. Take the very definition of "Islam." From Karen Armstrong to Bernard Lewis--and that's a pretty broad range--virtually every scholar of note (and many who aren't) has translated the term "Islam" as "submission to God." But "submission" evidently sounds off-putting to the American ear, so Esposito and Mogahed offer a different, more melodious translation--"a strong commitment to God"--that has a ring to it of everything but accuracy.

Or take the authors' cavalier attitude to the word "many." How many is many? Thirty percent of the vote won't get Hillary Clinton nominated for president, but it would be a lot if the subject were how many Americans cheat on their taxes or beat their wives. At the very least, one might expect a book based on polling data to be filled with numbers. This one isn't. Instead, page after page of Who Speaks for Islam? contains such useless and unsourced references as "many respondents cite" this or "many Muslims see" that.

Or take the authors' apparent indifference to facts. Twice, for example, they cite as convincing evidence for their argument poll data from "the ten most populous majority Muslim countries," which they then list as including Jordan and Lebanon, tiny states that don't even rank in the top 25 of Muslim majority countries. Twice they say their 10 specially polled countries collectively comprise 80 percent of the world Muslim population; in fact, the figure is barely 60 percent.

These problems would not matter much if the book gave readers the opportunity to review the poll data on which Esposito and Mogahed base their judgments. Alas, that is not the case. Neither the text nor the appendix includes the full data to a single question from any survey taken by Gallup over the entire six-year period of its World Poll initiative. We, the readers, either have to pay more than $20,000 to Gallup to gain access to its proprietary research or have to rely on the good faith of the authors.

Or, more accurately, we have to rely on Gallup's good name--the "integrity, trust and independence" cited above. Public comments by Mogahed at a luncheon I hosted at the Washington Institute on April 17 show exactly what that is worth.

Here's the context: As the event was about to close, Mogahed was pressed to explain the book's central claim that radicals constitute 7 percent of the world's Muslim population. A questioner focused on the critical distinction between the 7 percent of respondents who said the 9/11 attacks were "completely justified" and the other 93 percent. How many of those 93 percent, Mogahed was asked, actually answered that the attacks were "partly," "somewhat," or even "largely" justified? Were those people truly moderates?

In her answer, transcribed below, Mogahed refers in pollster code to numbers ascribed to the five possible answers to the poll question about justifying 9/11. Although she and Esposito never discuss the details of this question in their book, they did expound on them in a 2006 article in Foreign Policy magazine, which described a five-point scale in which "Ones" are respondents who said 9/11 was "totally unjustified" and "Fives" those who said the attacks were "completely justified."

In that article, she and Esposito wrote: "Respondents who said 9/11 was justified (4 or 5 on the same scale) are classified as radical." In the book they wrote two years later, they redefined "radical" to comprise a much smaller group--only the Fives. But in her luncheon remarks, Mogahed admitted that many of the "moderates" she and Esposito celebrated really aren't so moderate after all.

MOGAHED: I can't off the top of my head [recall the data], but we are going to be putting some of those findings in our [updated] book and our website.

To clarify a couple of things about the book--the book is not a hard-covered polling report. The book is a book about the modern Muslim world that used its polling to inform its analysis. So that's important: It's meant for a general audience, and it's not meant to be a polling report. One very important reason why is because Gallup is selling subscriptions to its data. We are a for-profit company; we are not Pew. We are Gallup. So this isn't about ... it was not meant for the data to be free since we paid $20 million to collect [the data] ... that we paid all on our own. So just to clarify that...

So, how did we come up with the word "politically radicalized" that we unfortunately used in the book? Here's why: because people who were Fives, people who said 9/11 was justified, looked distinctly different from the Fours. ...At first, before we had enough data to do sort of a cluster analysis, we lumped the Fours and Fives together because that was our best judgment.

QUESTIONER: And what percent was that?

MOGAHED: I seriously don't remember but I think it was in the range of 7 to 8 percent [actually, 6.5 percent].

QUESTIONER: So it's seven Fours and seven Fives?

MOGAHED: Yes, we lumped these two and did our analysis. When we had enough data to really see when things broke away, here's what we found: Fives looked very different from the Fours, and Ones through Fours looked similar. [Mogahed then explained that, on another question, concerning suicide bombing, respondents who said 9/11 was only partially justified clustered with those who said it wasn't justified at all.] And so the Fives looked very different; they broke, they clustered away, and Ones through Fours clustered together. And that is how we decided to break them apart and decided how we were to define "politically radicalized" for our research.

Yes, we can say that a Four is not that moderate ... I don't know. ...You are writing a book, you are trying to come up with terminology people can understand. ...You know, maybe it wasn't the most technically accurate way of doing this, but this is how we made our cluster-based analysis.

So, there it is--the smoking gun. Mogahed publicly admitted they knew certain people weren't moderates but they still termed them so. She and Esposito cooked the books and dumbed down the text. Apparently, by the authors' own test, there are not 91 million radicals in Muslim societies but almost twice that number. They must have shrieked in horror to find their original estimate on the high side of assessments made by scholars, such as Daniel Pipes, whom Esposito routinely denounces as Islamophobes. To paraphrase Mogahed, maybe it wasn't the most technically accurate way of doing this, but their neat solution seems to have been to redefine 78 million people off the rolls of radicals.

The cover-up is even worse. The full data from the 9/11 question show that, in addition to the 13.5 percent, there is another 23.1 percent of respondents--300 million Muslims--who told pollsters the attacks were in some way justified. Esposito and Mogahed don't utter a word about the vast sea of intolerance in which the radicals operate.

And then there is the more fundamental fraud of using the 9/11 question as the measure of "who is a radical." Amazing as it sounds, according to Esposito and Mogahed, the proper term for a Muslim who hates America, wants to impose Sharia law, supports suicide bombing, and opposes equal rights for women but does not "completely" justify 9/11 is ... "moderate."